Cadence Design Systems, Inc. (CDNS) Earnings Call Transcript & Summary

Good morning, everybody. Thanks for joining us, and congratulations on making it to the third and final day of the Wells Fargo Securities TMT Conference here in Las Vegas. My name is Gary Mobley. I'm one of the semiconductor analysts here at Wells Fargo Securities. And with us today, we have Nimish Modi from Cadence Design. Before I introduce him further and try to give you his title. I'm just going to hand it over to him, and he can give you a little background on himself and the company and as well read the safe harbor. So Nimish, thanks for joining us.

Thanks, Gary. Good morning, everyone. So today's discussion will contain forward-looking statements and could make use of certain non-GAAP financial measures. Please see Cadence's most recent 10-K, 10-Q and website for a discussion of risk factors and our use of non-GAAP financial measures. So I've been at Cadence, just a little bit of background over here about me. I've been in Cadence for about 16 years, and I run the Strategy & New Ventures Group. And my responsibility is our corporate strategy, corporate and strategic marketing, mergers and acquisitions, business development and recently the Molecular Sciences Group, which includes our OpenEye acquisition. In my earlier days at Cadence, I was in R&D and ran the front-end group and the system and software group. And then prior to Cadence, I was at Intel for 18 years and most recently in charge of the server CPU R&D, which includes the Itanium and the Xeon product families.

Thank you for introducing that, introducing yourself to us rather. Can you give us kind of an overview of how Cadence has evolved? I think everybody in the room and online knows specifically what you do on the EDA front, but maybe how you have moved into adjacencies and diversified the business more recently.

Yes. So Cadence, we are executing to our Intelligent System Design strategy. And when you think about the strategy at a high level, the foundational aspects of the strategy are our core EDA and IP business. And we provide on the core EDA front a full end-to-end flow across analog custom digital design and verification. And then we focus on a profitable and scalable IP business, in specific verticals are the most advanced nodes. And then to your point, we have looked at expanding or extending our computational software, legacy and heritage and strength in computational software to expand into new adjacencies with system analysis, system innovation being the next one. So this is all about system design and analysis, a PCB design, packaging design and more recently into multiphysics and molecular modeling with OpenEye. And then the third layer is all about data, and we call it the Pervasive Intelligence layer. So when you think about this in the context, for example, of our complete system and automobile, for example, you've got the data piece, which is all tied into the navigation and the like. And then you got the actual system, which is the car itself, the chassis, and then you got the chips, underlying chips, which are basically driving the car. So that's -- we service all these 3 aspects. And so it's really, as we look beyond core EDA and IP, we think that there's a great TAM expansion opportunities. We've talked about tripling our TAM from $10 billion TAM to $30 billion over 5 years, and we are rolling our way in that expansion.

Yes. I suppose since you and your main competitor have such dominant market share in EDA, the way to grow is moving into adjacencies, although the EDA market certainly has hit a renaissance in terms of growth. And related to that, I want to hit on what appears to be an emerging important product cycle on the EDA front, and that is AI and machine learning and Jed AI or JedAI...

JedAI.

Is the underlying technology for a different array of products, Cerebrus, Verisium and perhaps soon some additional products. Maybe if you can give us an overview on the strategy there and why these tools are so transformative to the industry.

Yes. Gary, this is very exciting times in that regard. I mean, we are super excited about all these new AI-driven tools that we have come up with. You mentioned Cadence Cerebrus optimality on the systems for verification. And JedAI is actually our big data analytics platform, which kind of connects all of these 3 different solutions and the results are pretty transformative. But before I get into that, just when you take a step back and look into it from a design flow perspective, EDA has always really focused on a single run if you will. So through designer intrusion, design experience, it's all focused on coming up with algorithms, which will optimize the output of one run. And then the designer just iterates through it to converge. There's never been a mathematical way, if you will, to take the output of one run and feed it to the next run. So what we've done is to provide a mathematical and algorithmic framework by means of which you transfer the learnings from one run to the next and do that in an automated manner. So that really helps with the conversions much faster. So that's one aspect of it. But it's not just about conversions. When folks are doing designers are doing simulation and the like, in the end, the goal is to have a better product. It's all about optimization as well. And so this is where it comes in. And so we have developed some very sophisticated AI algorithms, which will intelligently explore the design space and then uses reinforcement learning to come up with better PPA, power performance area, and it doesn't really require a lot of a priority training. It actually trains on the fly if you will, kind of cold start. So it's really early days still, but the results our customers have been seeing have been just amazing. I mean we are seeing up to 30% improvement in quality of results of PPA. When you think about this, this is like on a similar scale to some process node -- the benefits that you get from some process nodes migrations. We've got 10 of the top 20 semiconductor companies using Cadence Cerebrus, for example, over 200 production designs. And several of these customers have talked about their successes over 20 papers at our Cadence Live conferences this year. And Optimality and Verisium are more recent, but the ramp is happening very well as well. So when you look at it from a customer's perspective, design complexity is kind of exponentially growing and there's a growing shortage of talent of experienced engineers. So what solutions like Cadence Cerebrus and the like do is, they kind of democratize the chip and system development cycle. You think of them as like engineering assist tools, which are very far reaching in terms of their impact for both productivity and quality of results. And so yes, its early days, but the results speak for themselves, and we are very happy with this. And I think the key over here is we started off with Cadence Cerebrus, which was on digital, but then we -- as I said, we moved on and expanded beyond that into verification into the systems space and PCB and the like.

Okay. Well, I mean, to summarize, in simple terms, it sounds like it subsidizes your customers required engineering man hours and leads to a better chip design.

It leads to a better chip, yes.

And so it's a win-win, presumably for everybody involved. So from your perspective, how does it translate into a higher average deal size? Or said differently, EDA vendors like yourself grabbing more of the R&D wallet share?

Yes. So there's a lot of value that is being provided over here by us, and I think we fully expect to benefit from that value as more of the spend of customers starts moving away from headcount to these very sophisticated automation solutions. So we will see that progressively moving. The other aspect of this is while we monetize these solutions directly, I mean, they are built on top of our existing products. So I mean, the underpinning these solutions are the existing products. So there's an indirect pull-through effect of these underlying products and makes them sticky and also we get many more fulfill wins as a result. And so yes, like I said, it's early days, but we are doing our best to working with customers on the discovery and aligning our -- the pricing to match the value that we are providing. And -- but in the end results, they speak for themselves. And I think that, in turn, kind of ends up leading to much more productive discussions on that front.

Got it. I know you touched on this in your opening remarks, but maybe if you can give us an overview of what you're trying to accomplish in the system innovation and systems analysis segment of the market, specifically multi-physic simulation computation dynamics in your recent acquisition of OpenEye.

Sure. So when you look at the electronics ecosystem, there's a lot of hyper convergence happening of mechanical and electrical hardware and software, semiconductors and systems. And all of this requires these design tasks to be done concurrently. So there's a lot of co-design co-verification, costimulation, which is required. And also stating the obvious, but there's just increased pressures on system time to market. But what all of that is leading to is a shift left paradigm that we've talked about in the past, that's just growing momentum on that front, meaning you want to simulate as much as you can as early as you can so that you can flush out all these issues before you commit to silicon or commit to manufacturing. So what we've done is we have extended our heritage or lineage in EDA. And we've got a very deep-rooted expertise in chip level simulation and extended that to the system level, so system-level simulation or system analysis. And I mean, system analysis market by itself is a very attractive TAM expansion opportunity for us, growth opportunity for us. We are talking about $7 billion or $8 billion market growing at double digits. And then the financial profile of that business is very aligned to our operating model as well in a steady state. We fully expect the margins of that business to be equal, if not better than core EDA. And so the vision over here is just like what we have done in other parts of our business is to develop best-in-class tools and then provide a comprehensive platform, a comprehensive multiphysics platform. We started off with finite element analysis with electromagnetics and electrothermal and then expanded into computational fluid dynamics. And we brought a full flow, right, from the preprocessing, meshing, solvers, optimization post processing. And so that's basically what the strategy we're here is. And the customers have been asking us to move into system design and analysis because they want their solutions to take a holistic system level view. And system design and analysis market, system analysis is a part of system design and analysis is about 12% of our business right now, and we've been growing at about 30% for the last few quarters. Now on to your other question on OpenEye. And -- so similar mindset or thought process over here, we want to extend our computational software expertise to expand into newer areas. So we've done that with on the multiphysics front, as we talked about. And over there, the target end markets, high-tech electronics, aerospace, defense, automotive. Now when you look at further down the road as we build up the multiphysics space, we thought there was a great opportunity to apply that same expertise that we bring to the table to biosimulation, life sciences market, specifically for drug discovery. And when you look at that market, it's physics based. I mean we are very good at transistor modeling and simulation. This is molecular modeling and simulation. So we can bring a lot of that expertise, capabilities we are very adept at to bring some really disruptive innovation to that space. And AI, I mean as you look at AI-driven drug discovery and the like, I mean, these are all future areas of innovation, and they align very well with all the things we talked about in the AI front that we are doing with our pervasive intelligence strategy. So yes, this is the multiphysics aspect of it, we're building in -- up into it. It's the here and now and for the next few years. And then as you look a little bit further down the road, this is where we think the biosimulation space is really going to be. A very attractive opportunity for us.

Thanks for that. One of the common questions that I get from investors is trying to get a better understanding of why it is right for you to move into the multiphysics simulation software world. What makes your approach or your technology better than some of the more established players that are already in the market.

Yes. So some of it is exactly in the lines of what we talked about. I mean that's a good question, right? When you're entering a new space, what's the new value that you bring that, right? What's your right to win? And the computational software expertise, I think that's the key aspect over here that we are very adept at the computational software. When you look at this over the years, we have developed some very sophisticated algorithms, very sophisticated solvers to address all these very complex chip design challenges. And these are an order of magnitude, more complex than solids than physical simulations. And also, when you look at the rate of innovation in the chip design space, it's really driven by Moore's Law. Every 2 to 3 years, you've got to come up with a new innovative -- new wave of innovation to meet the technical challenges of that next wave and take advantage of those opportunities. And we've innovated on massively parallel architectures, distributor processing, very sophisticated algorithms, all to improve performance of the chip without really compromising accuracy. And that's basically what we bring to the table over here to the multiphysics space. And the results speak for themselves. I mean, the first 2 products that we came up with, Clarity Electromagnetic Solver and the Celsius Electrothermal Solver were organically built. And the customers are seeing up to 10x performance improvement over legacy solvers without compromising accuracy. I mean -- and so right there, you can see the benefits there. And then, of course, we moved on into -- further into competition for dynamics and the like. So I think this is a very -- as I said, a very attractive growth opportunity for us. And most importantly, we feel like we've got a really good value proposition that we bring to it and the results speak for themselves.

Okay. I want to shift the conversation to another hot topic as it relates to the EDA industry, and that's China. And so maybe if you can give us an overview of the mix of revenue that you received from China and the more I guess, appropriate what the mix of revenue is from China indigenous customers. And the reason I'm bringing it up is, of course, the U.S. export controls and then we'll double-click from there.

Yes. So I think -- it's just worth restating, I think we've got a very resilient business model, and we are very diversified, both in terms of our customer base as well as regionally. And then China is one of the many global markets that we have. As far as the U.S. export control restrictions go, I mean, in our recent Q3 earnings call, we stated that we've studied the recent -- the newly announced export control restrictions, and we believe the impact to be limited and manageable. And of course, we are in compliance with all the U.S. export control regulations. Now to your question about the revenue mix over there, we get our business, or we do business with both the multinational companies as well as domestic Chinese companies. Now we don't provide the details there, but the majority of our business there is with the domestic Chinese companies over there.

Okay. And your main competitor last night had something similar to say about managing through these export controls manageable, I guess, is phrasing what they had to say as well. But I wanted to maybe get your view based on what the U.S. commerce department is trying to accomplish in these export controls and kind of help us understand sort of a worst-case scenario if this is, in fact, death by 1,000 cuts as it relates to serving those indigenous Chinese customers. I guess asked differently, what portion of your business is exposed to digital implementation in those areas that seem to be a focal point for the U.S. commerce department?

I mean, so they have been progressively over the last few years, the export control restrictions. There are many components to that. But the most recent ones, the 3D-6, [indiscernible] Rule for example, that is tied into digital implementation and specifically at the most advanced nodes. And what we've done is most of our tools are horizontal in the sense that they're not really tied to any one technology node per se, but in this case, we have created a gap free version of the digital implementation tools that we can continue selling in there as well. And so yes, as we said, I mean, we've studied the impact of regulations, so we believe that to be limited and manageable not just for the guidance that we have provided for 2022, but we believe that's to be the case for 20 -- fiscal year '23 as well, even though we are not providing guidance for '23 at this time.

Okay. Well, I guess on the opposite, I know that's topic to speak about the opportunity in China, in the capital that seems to be flowing quite freely there. You have a number of hyperscalers over in China, Alibaba and Baidu, specifically a number of automotive OEMs like BYD. They've initiated plans to IPO their chip ARMs. And so I'm wondering perhaps if that's beneficial to you because of better access to capital or for focus.

Yes. I mean, to me, this is just a sign. They're raising more capital because they probably want to invest more into service the growing demand that they see from their chip ARMs. And this is kind of reflective of a broader global trend of more and more silicon activity happening. We have talked about generational trends in our earnings calls before, 5G, hyperscale, Autonomous, AI/ML. I mean these are all mega trends which are playing out over many, many, many years, and each one of them is in a different stage of its life cycle. And that's why we call them generational trends. I mean they're truly not a short-term fad. I mean they're just going to play out over multiple years. And they're all driving tremendous amount of silicon activity. And so to us, this is just reflective of more silicon activity, more silicon activity is more design activity, more design activities, more design starts, more design starts, it's all good for EDA.

One of the other questions that I get is how the softening chip market might impact the licensing environment for EDA. I understand that you have a long duration of revenue recognition model, typically 2.7 years of contract life. But if bookings soften because of a soft chip environment, your backlog course is going to decline. And so I'm curious to know -- hear your opinion on how R&D activity might be affected by the softening chip market, if at all?

Yes. So while there may be a softening of the chip cycle of the semi cycle, again, it's very important to understand that design activity, EDA is all predicated on semiconductor R&D. And that's the first order to connect it to semiconductor R&D spend versus manufacturing or the volume of design silicon sold, if you will, by our customers. And the semiconductor R&D budgets, if you will, design activities, depends on the type of chip, which is being designed, but typically design activities for those chips precede 18, 24, 30 months before the chip ends up in production. And so the design activity of today are going to see manifest themselves in silicon 1.5 years down the road or so. So typically, in these kind of scenarios, situations, what we have seen is customers, especially the larger ones, they actually view these guys in the situations as an opportunity to create further separation from a competitive standpoint and they actually invest more because they want to come up with a stronger road map and stronger products on the other side of the downturn. The thing which you don't want to happen is when you come out of the downturn that you don't really have a strong product out there. And so far, we continue seeing strong robust design activity and...

Great. We talked about this when we had a chance to have breakfast this morning, but the -- it's clear that more system OEMs and hyperscalers are taking on their own chip development efforts for various different reasons. I mentioned that I counted somewhere close to 40 different system OEMs developing their own ships internally. If you go back 10 years ago, you can probably count that number on one hand. And so I'm hoping to hear from you how your business is transitioned to be more centralized around system OEMs or shifting towards system OEMs.

Yes, Gary. I mean this is really very exciting, right? Because we are seeing more and more system companies investing in their own custom silicon. I mean it's really their own secret sauce, differentiated IP that they want to basically optimize for the targeted workloads and do that in a very much cost-efficient manner, risk management and the like that we talked about. Now as -- if you look over the past few years, as our portfolio has both expanded and proliferated, we've done very well with both the semiconductor companies as well as the system companies. I mean the tide has lifted both boats for us. But I would say that the ramp up or the slope, if you will, of the system companies is probably more pronounced and about 45% recently -- about 45% of our revenue does come from system companies. And the other thing also is that when you guys are system companies as customers, it's not just about the core EDA and IP. There's also an opportunity for other components of our system stack that we provide, packaging, PCB, system simulation to open up doors for that as well. And yes, and this trend continues.

It's hard to imagine that this redundancy of chip design can be sustainable. What I mean by that is Apple has chosen to develop many of its own chips for its different products. And that comes at the expense of the traditional merchant chip companies. So my question to you is, how does this redundancy of chip design become accretive to the EDA -- the growth of the EDA industry long term?

Yes. So a little bit of a correction, if you will, over there. We don't see this as redundancy. We think it's actually additive to the TAM or accretive to the TAM. And let me explain why. So when a system company is building its custom silicon, so prior to them building custom silicon, they were purchasing merchant silicon, right? And now when they start, when the system companies start building their own silicon, their merchant silicon vendor loses that volume because it shifted over to the system company, but the merchant silicon vendor still has to design the chip because they've got other customers for that, right? So all the design activity, which was happening on the merchant silicon side is still continuing, which is, again, we are predicated not by volume of shipments, but by design activity, that's continuing. And on top of that now, you've got new design starts because the system company is now building their own custom silicon. So it's really a digit to that. And this trend, again, you go back to the reason why folks are doing this all the multiple reasons why they are doing this. I think these -- this trend is going to continue. I think system companies are going to invest more in building custom silicon and that trend is going to continue and proliferate into other verticals as well. So I think this is -- as I said, it's additive and it's going to be value accretive to our business as it already is.

Okay. Cadence in the EDA market is always focused on -- primarily on the analog IC design community. I mean I think arguably, that's definitely been your strength, but you certainly have a good digital IC design tool set as well. Can you speak to why it's easier for you to take share in the digital EDA market realm versus conversely your competitor coming over and taking share in the analog. Why is the analog IC design community stickier for you as a customer base?

Yes. So I mean on the digital front, we have made giant strides, we're taking giant strides over there in the last several years, right? Because we have revamped our entire portfolio. You come up with new algorithms, new solvers, massively parallel architecture distributor processing. And it's not just about coming up with leadership best-in-class tools there, but the full flow as well. And if you look at the results, I mean, a few years ago, of the top 20 semiconductor companies, we probably had a digital footprint in 2 or 3 of them. And now we are in all 20 of them. I mean -- and so I think that shows just how we have progressed over time there. Now when you think about this in the context of digital or analog and digital, and as you pointed out, we've had a very long-standing entrenchment of our Virtuous platform, analog platform for the last 20-plus years. Digital is more about, when you think about this is a propagation more of binary values, right, 0 and 1. And analog modeling is different, right? That's more of a continuous time domain kind of behavior that you're modeling. And the signal integrity performance become very important. And you've got to compensate -- the tools need to compensate for any variability, which is happening in environmental conditions when you're sensing something or manufacturing variability. I mean it's very complex design challenges, which the tools need to address. I mean, analog design is as much art as a science, if you will, and analog designers, when they've got a flow and tools and processes and methodologies which work, they tend to stick with them. And increasingly, right now, it's not just about single solutions over there and coming up with an analog chip. You got to look at it from a system level as well. So it's about coming up with the solutions which are addressing analog, mixed signal, RF packaging. And packaging is becoming, as you know, extremely important advanced 3D packaging. And I think just given where we are with our Virtuoso platform and some of the other augmentations that we have done around it with the flow in a really good position over there.

Okay. Clearly, Cadence has had a really good fiscal year '22. I know the year is not finished, but per the guidance, it's going to be a really good year, much better than what you expected when you started the year. But I think much of the upside has been driven by upfront revenue, which largely includes verification hardware. It's been a great product cycle for you on 2 different fronts. And so maybe if you can speak to why it surprised the upside throughout the year? Where you sit with respect to backlog in that particular segment? And how long do you think this product cycle strength may continue?

Yes. So verification, I mean continues to be one of the biggest challenges customers face, right? I mean, verification is never done. It's never done. I mean it's just more a question of how much verification you do to raise your confidence before you tape out or before you commit to silicon or manufacturing. And we've got a full flow verification platform with both software products and hardware products and software simulation, formal analysis. And on hardware, we've got emulation and prototyping. And so when you look at it from that perspective, the complexity of design is just out scale, the ability to do simulation at any reasonable level -- with any reasonable level of confidence. And so emulation is just a must-have part of the flow. I mean, that dynamic has changed where it was more discretionary, if you will, a few years ago, and you tried to squeeze in as many cycles as you can before you take out versus now saying, if you haven't done this much, you're heading for trouble. I mean, so it's just a must-have part of the flow. And it's not just chip level verification, it's system level of verification. And then there's a software bring-up piece of it as well, which is why the prototyping platforms come in. So our Palladium Z2 and Protium X2, what we call the Dynamic Dual platforms, done exceedingly well. And it's just a -- we call it a secular demand. And I think that's something which, from a demand perspective and the need is going to continue. Now it's a lumpy business. It's more project driven. And because of the upfront revenue, it kind of tends to be cyclical in that -- which is why we encourage folks not to look at 1 quarter or 1 half. But you look at where the trends are, where they're going. and these platforms are extremely successful with customers and the share of common front-end compiler, which makes it very easy for folks who are already using Palladium to kind of just use Protium and bring up the platform very quickly. So this year, we've had -- it's been the best year that we've had on hardware. And we have seen a lot of interest, not just from existing customers, but also from some of the newer one. Hyperscalers have been a big part of EDA, companies are a big part of that as well. And so stay tuned.

All right. Nimish, we're out of time. I appreciate you joining us today, and I appreciate all the people in the audience join us as well and as well as those people online. Thanks, again.

Thank you, Gary. Thanks, everyone.